News Stay informed about the latest enterprise technology news and product updates.

Master data management must start with data governance

An effective master data management project starts with managing people and processes, including data governance. The technology is secondary.

WALTHAM, Mass. -- Executives at Cambridge, Mass.-based Cambridge Interactive Development Corporation (CIDC) can't...

even agree what a download is. So before beginning its master data management (MDM) project, the online gaming software company is first setting out to standardize metadata definitions.

CIDC's dilemma is a common one, according to Rick Sherman, data management consultant and founder of Athena IT Solutions. Speaking last week at The Data Warehousing Institute's Boston chapter meeting, held at Children's Hospital Boston in Waltham, Sherman told attendees: "It's easy to define what master data is. The trick is the management part of master data management."

Agreeing on metadata definitions, such as what constitutes a product, and articulating the business rules around master data are critical to a successful master data management project, Sherman told attendees. Dealing with the personalities involved and managing expectations is no small task either. The technology is secondary.

"If tools could do this for you, you would pay any price to get those tools," Sherman said. "The tough stuff is the 'P' stuff: the people, politics and processes. We'd all like some toy to fix our data problems, but it's not going to happen."

Data governance the first step in master data management for a software company

Meeting attendee Alex Chervinsky is a business intelligence developer/database administrator with CIDC. He said the software company is close to finalizing a deal with San Francisco-based Data Advantage Group to use its MetaCenter MDM tool to consolidate business and operational data.

The impetus for CIDC's MDM initiative is a lack of data consistency among its various department-level reports, which sometimes results in confusion on the part of managers and others, Chervinsky said. For example, some at CIDC define a download as just that: when a customer simply downloads CIDC software from the Internet. Others consider a download to have occurred only after the customer actually installs the software. CIDC executives recognized that without standardizing metadata definitions before an implementation -- just as Sherman suggested in his presentation -- any MDM initiative wasn't likely to reduce the confusion.

"How are you going to create a metadata [repository] if you haven't had the interviews with people to agree on the definitions?" asked Sean Knight, a data manager recently hired to lead CIDC's metadata definition project. "So we're trying as hard as we can to start some of the functional discussions with people so that [metadata definitions are] not going to become a roadblock when our technical environment is up and running."

More on master data management and data governance
Find out whether master data management is making headway in the public sector  

Learn how integration competency centers can help centralize data integration projects  

Check out how one customer data integration tool's fuzzy search helps clear up call center confusion

Knight first gathered the 100 or so reports deemed most critical by CIDC executives and managers. He then identified the key metrics within those reports that needed agreed-upon definitions. The next step, which Knight plans to begin soon, is to make some "intelligent guesses" as to what those metric definitions should be, then to send them to a group of about five or six representatives from CIDC's various departments.

The departmental liaisons will supply feedback, at which point Knight will evaluate and identify any metadata definition discrepancies. Then CIDC can get to the difficult work of actually reconciling the discrepancies and agreeing upon definitions.

Knight hopes eventually to create a wiki-style company glossary of metadata definitions. "We want to make everybody responsible for conforming to the way we've all agreed to define things," he said, "but we can't expect to do that if they don't know how things are defined."

While Knight continues reconciling CIDC's metadata definitions, Chervinsky will begin laying the groundwork for the MDM implementation. When both tasks are completed and the MDM project goes live, the result, both men agreed, will be quicker access to more accurate, cleaner data for CIDC marketing executives, sales representatives and other personnel.

Is your company on the verge of implementing an MDM tool? Email editors with your story!


Dig Deeper on MDM best practices

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.